Breast cancer, like other common cancers, shows familial clustering. Numerous epidemiological studies have demonstrated that, overall, the disease is approximately twice as common in first degree relatives of breast cancer patients1. Family studies, and particularly twin studies, suggest that most if not all of this clustering has a genetic basis2,3. For example, Peto and Mack3 estimated that the risk of breast cancer in the MZ twin of an affected woman was approximately four-fold greater than the risk to a sister of a case.
Several breast cancer susceptibility genes have already been identified, most importantly BRCA1 and BRCA2. Mutations in these genes confer a high risk of breast cancer (of the order of 65% and 45%, respectively, by age 70)4. Mutation screening of population-based series of breast cancer cases has shown that only about 15% of the familial risk of breast cancer can be explained by mutations in these genes5,6. The other known breast cancer susceptibility genes (TP53, PTEN, ATM, CHEK2) make only small contributions to the familial risk (because the predisposing mutations are rare and/or confer only small risks). In total therefore, the known breast cancer susceptibility genes have been estimated to account for no more than 20% of the familial risk7.
Genetic variation in risk may result from rare highly-penetrant mutations (such as those in BRCA1 and BRCA2) or from variants conferring more moderate risks. Several lines evidence suggest strongly that high penetrance mutations are not major contributors to the residual familial risk of breast cancer. Firstly, mutation screening of multiple case families has found that the large majority of cases with a very strong family history (for example four or more affected relatives) harbor mutations in BRCA1 or BRCA28. Secondly, despite extensive efforts over the past nine years, genetic linkage studies have not identified any further linked loci9,10. Thirdly, segregation analyses of large series of breast cancer families have found, after adjusting for BRCA1 and BRCA2, no evidence for a further major dominant breast cancer susceptibility allele11,12. In the largest such analysis, Antoniou et al.13 found that the most parsimonious model for breast cancer was a polygenic model, equivalent to a large number of loci of small effect combining multiplicatively.
While the above analyses suggest, that several low penetrance breast cancer susceptibility genes remain to be detected, the precise number of such genes is unknown. Moreover, in the prior art, it is unclear whether such susceptibility alleles are common or rare in the population. The subject application focuses on alleles that are relatively common (frequencies greater than 5%) and identification of such loci is performed herein on a genome-wide basis.